Funso Popoola

I implemented dataflow pipelines that process around 13.4GB/hr of streamed data into a warehouse that provides a basis for customer segment analysis and wrote SDKs that made the analytics solution extensible to many airline clients with the minimum configuration.

I developed a code for a Python program that enables the seamless conversion of English textual numbers into Arabic numerals (figures), such as 'two thousand and seventeen,' into the corresponding numerical representation, in this case, 2017.

To run this code and experience its capabilities, simply execute the 'translator.py' file using a Python 2.7.x interpreter through the terminal.

At OjaExpress, I designed and implemented an eCommerce platform that delivers ethnic groceries to Africans and Caribbeans in the US.

While working with Commutelab (a UK startup), I designed and implemented a software to convert UK transport data from TransXchange XML format to Google Transit Feed Scheme (GTFS) format.

The UK transport data is released fortnightly and needs to be converted to GTFS upon every release for the company's use of the services it provides to its clients. Hence, there were limits on the storage space used and the processing time and memory. We were able to optimize the implementation of this software to process the gigabytes of data provided in few hours while running on a machine with an average memory of 2GB RAM.

This software was written purely in Python, using a SQLite database for storing intermediate results.

At Kangpe, I designed and implemented a suite of software components that allowed the company to provide a link between hundreds of verified professional medical practitioners and tens of thousands of users so they can obtain quick medical services while (mostly) preserving their anonymity.

I developed a suite of fast and scalable REST APIs that served the mobile and web clients used by both the medical experts and the other users.

Yorùbá numerals have been seen as one of the most interesting but quite complicated numeral systems. The system translates English numbers both in figure and text to its standard Yorùbá form. The computational processes underlying both numerals were used to formulate the model for the work. Unified Modeling Language (UML) and Automata theory was used for the system design and specification.

The designed system was implemented using Google Web App Engine with support for Python. The result of the system evaluation using the mean opinion score approach shows that the system gives a recall of 100% on all the output considered.

While wrapping up my undergraduate studies, I worked with a graduate student in implementing a system that used basic cosine similarity rule to grade student continuous assessments that are not the regular multiple choice tests.

It was a web application developed using the Flask framework and using SQLite as the storage technology.